A recent study showed that maternal treatment with zidovuding (ZDV) during pregnancy, labor, and infant treatment after birth resulted in a reduction in the transmission rate from 25% to 8%. This therapy was remarkable, but not 100% effective in a highly selected cohort of women. Even under ideal circumstances, 8% of infants were infected and in reality, many women do not receive early prenatal care and/or refuse ZDV therapy. Investigations of other therapies which may further reduce transmission, such as other antiretroviral drugs, HIVIG, vaccination, and vaginal treatments are underway. It must be anticipated that a finite number of infants born to HIV infected mothers will be infected even when the most effective preventative therapy is available. Effective therapies for these infants are needed. These infants represent a unique opportunity for intervention with gene therapy approaches at the earliest stages of disease. Introducing antiviral genes into primitive hematopoietic cells is a goal of most gene therapy approaches for AIDS, including our own involving HIV-regulated toxin gene (DT-A). Umbilical cord blood (CB) is a readily accessible source of primitive hematopoietic cells, which has been used for tranplantation in malignancies and immunodeficiency. However, the properties of CB from HIV-infected individuals have not been systematically explored. Demonstrating that sufficient quantities of progenitor cells can be expanded, genetically engineered to be HIV-resistant, banked for future use and transplanted is cruical for evaluating the ultimate feasibility of using engineered CB cells for transplantation into HIV-positive infants. We propose the following, using CD34+-selected CB from HIV-infected individuals: 1) evaluate CB cells from mothers stratified for CD4+ cell count for the percentage and absolute number of primitive, hematopoietic cells and the HIV-1 viral burden 2) determine the efficiency of retroviral transduction into CB cells using standard and pseudotyped viruses; (3) evaluate the anti-HIV effect of regulated toxin genes in CB cells; and 4) determine the characteristics of CB cells upon freezing and thawing, as the first step toward establishing a CB bank from HIV-infected individuals. We estimate that we will obtain 18-22 CB specimens from HIV-infected individuals per year. The sample size should be enough to provide the necessary information to evaluate whether transplantation with engineered HIV-resistant CB would be a feasible approach to preventing or treating AIDS in infants.